Multi-modal Emotion Recognition in Conversation (MMERC) aims to identify speakers' emotional states using multi-modal conversational data, significant for various domains.MMERC requires addressing emotional causes: contextual factors that influence emotions, alongside emotional evidence directly expressed in the target utterance.Existing methods primarily model general conversational dependencies, such as sequential utterance relationships or inter-speaker dynamics, but fall short in capturing diverse and detailed emotional causes, including emotional contagion, influences from others, and self-referenced or externally introduced events.To address these limitations, we propose the Evidence-Cause Attention Network for Multi-Modal Emotion Recognition in Conversation (ECERC).ECERC integrates emotional evidence with contextual causes through five stages: Evidence Gating extracts and refines emotional evidence across modalities; Cause Encoding captures causes from conversational context; Evidence-Cause Interaction uses attention to integrate evidence with diverse causes, generating rich candidate features for emotion inference; Feature Gating adaptively weights contributions of candidate features; and Emotion Classification classifies emotions.We evaluate ECERC on two widely used benchmark datasets, IEMOCAP and MELD.Experimental results show that ECERC achieves competitive performance in weighted F1-score and accuracy, demonstrating its effectiveness in MMERC 1 .

Existing studies on multi-modal emotion recognition in conversation (MMERC) mainly focus on multi-modal fusion and context modeling for emotion representation, facing limitations in uncovering the intrinsic structure of emotion-related data. The existing geometric consistency regularization (GCR) technique aims to build meaningful latent feature structures in multiple modalities and has been validated for non-conversational data, therefore we further explore its application to the MMERC task, which involves conversational data. However, we find that the geometric consistency in conversational data varies among different speaker and conversation relations (intra-speaker, inter-speaker, and inter-conversation relations), probably due to differences in speaker expressions and conversation topics. This makes the direct application of GCR less effective. To address this issue, we propose a Multi-Relational Geometric Regularization Framework for MMERC (R4-MMERC). Our framework constructs geometric structures of conversational data and performs dynamically balanced consistency regularization based on multiple speaker and conversation relations (intra-speaker, interspeaker, and inter-conversation relations). We tested our framework by integrating it with three typical MMERC models on the IEMOCAP benchmark dataset. The results show significant performance improvements, demonstrating the effectiveness of our approach1.

Solid State Drives (SSDs) based on the NVMe Zoned Namespaces (ZNS) interface can notably reduce the costs of address mapping, garbage collection, and over-provisioning by dividing the storage space into multiple zones for sequential writes and random reads. The Log-Structured Merge (LSM) tree, which is extensively used in key-value storage systems, converts random writes to sequential writes, hence a suitable scenario to utilize ZNS SSDs. However, LSM tree associated data significantly varies in lifetime due to the levels and merging mechanisms of the LSM tree. Therefore, without an accurate method to estimate data lifetime, data with disparate lifetimes may be placed in the same zone, thus causing low space utilization and high write amplification within the SSD. To address these issues, the article proposes two data overlapping aware optimizations to realize intelligent data placement: a zone allocation scheme and a garbage collection scheme. The key technique of these optimizations is an accurate data-lifetime estimation by considering both the associated tree level of the data and the data overlapping ratio between the data and those in the neighboring level. Using the estimation technique, the zone allocation optimization can place data with similar lifetimes in the same zone. Besides, the garbage collection optimization can reclaim zones in an adaptive manner based on overlapping ratios to reduce the amount of data migration. Experimental results demonstrate that the optimization schemes effectively reduce garbage collection-incurred data copy by average factors of 2.11× and 1.50× in comparison to a conventional work and a state-of-the-art work, respectively. Consequently, the proposed work successfully alleviates the write amplification effect by 18% and 6%, compared to the conventional work and the state-of-the-art work, respectively.

Lip reading uses lip images for visual speech recognition. Deep-learning-based lip reading has greatly improved performance in current datasets; however, most existing research ignores the significance of short-term temporal dependencies of lip-shape variations between adjacent frames, which leaves space for further improvement in feature extraction. This article presents a spatiotemporal feature fusion network (STDNet) that compensates for the deficiencies of current lip-reading approaches in short-term temporal dependency modeling. Specifically, to distinguish more similar and intricate content, STDNet adds a temporal feature extraction branch based on a 3D-CNN, which enhances the learning of dynamic lip movements in adjacent frames while not affecting spatial feature extraction. In particular, we designed a local–temporal block, which aggregates interframe differences, strengthening the relationship between various local lip regions through multiscale convolution. We incorporated the squeeze-and-excitation mechanism into the Global-Temporal Block, which processes a single frame as an independent unitto learn temporal variations across the entire lip region more effectively. Furthermore, attention pooling was introduced to highlight meaningful frames containing key semantic information for the target word. Experimental results demonstrated STDNet's superior performance on the LRW and LRW-1000, achieving word-level recognition accuracies of 90.2% and 53.56%, respectively. Extensive ablation experiments verified the rationality and effectiveness of its modules. The proposed model effectively addresses short-term temporal dependency limitations in lip reading, and improves the temporal robustness of the model against variable-length sequences. These advancements validate the importance of explicit short-term dynamics modeling for practical lip-reading systems.

NVMe Zoned Namespace (ZNS) devices partition the storage space into sequential-write zones, notably reducing the costs of address mapping, garbage collection (GC), and overprovisioning. Log-Structured Merge (LSM) tree-based databases convert random writes into sequential writes and can thus be efficiently handled by ZNS devices. Efficient zone-allocation methods play a pivotal role in maximizing the performance of LSM tree-based store running on ZNS devices. However, existing zone-allocation methods encounter high write-amplification factors due to inaccurate lifetime estimation solely based on the LSM-tree levels. To address this, this paper proposes an overlapping-aware zone-allocation method, termed OAZA, which efficiently selects suitable zones to place data. First, OAZA estimates the data lifetime by considering both the LSM-tree level of the data and the relative data hotness within the same tree level. Secondly, OAZA intelligently selects an appropriate zone to store the data based on the estimated lifetime. Experimental results demonstrate that OAZA outperforms two zone-allocation methods that correlate data lifetime merely to the tree level. Specially, OAZA reduces the amount of GC-induced data copy by average factors of 2.7 × and 1.7× in comparison to the two methods, respectively. Additionally, OAZA achieves an impressively low write-amplification factor of 1.1 ×, outperforming the factors of 1.2× and 1.3× achieved by the two compared methods, respectively.

The emerging Zoned Namespace (ZNS) interface helps flash-based SSDs achieve high performance by dividing the logical space into fixed-size zones. Typically, a zone is mapped to blocks across multiple dies to achieve I/O parallelism. Small zones can make better use of space and are therefore widely studied. However, a small zone fails to be mapped to blocks residing on all dies, causing underutilized die-level parallelism. Meanwhile, a fine-grained (i.e., plane-level) parallelism is rarely exploited for ZNS SSDs due to a strict limitation mandating that only the same type of operation can be simultaneously performed on the same address across different planes within a die. To address these issues, this paper proposes a novel small-zone ZNS-SSD design with dynamic zone mapping, named Para-ZNS. First, a new parallel block grouping module is devised to group blocks across all planes from multiple dies as a basic unit to be mapped to a zone. Such a basic mapping unit achieves parallelism among multiple dies and plane-level parallelism. Then, a die-parallelism identification module is implemented to locate idle dies. Subsequently, to fully exploit the die-level parallelism, a dynamic zone mapping scheme is employed to intelligently map the basic mapping units on the identified idle dies to open zones. The evaluation results based on a widely-used I/O tester (FIO) demonstrate that Para-ZNS improves the bandwidth by 3.42× on average in comparison to state-of-the-art work.

This paper proposes a rumor control model based on community immunization. Based on the community division and the trust network inference algorithm, the model redefines the standard to measure the importance of nodes in the network. First, the model uses the Louvain clustering algorithm based on the Ochiai coefficient to discover the network community and then presents the trust network inference algorithm. By analyzing the key factors that affect trust transfer between nodes, the trust evaluation between unfamiliar nodes is inferred, and important nodes with a high degree of trust in the network community are calculated. Finally, combined with the characteristics of inner degree and outer degree centrality of nodes in the network community, five types of important nodes in the network are screened out. To avoid repeated selection of nodes, this paper identifies a group of key nodes in the network community for local immunization by means of deduplication and taking intersection, so as to realize effective control of rumors in the network.

Two‐party private set intersection (PSI) plays a pivotal role in secure two‐party computation protocols. The communication cost in a PSI protocol is normally influenced by the sizes of the participating parties. However, for parties with unbalanced sets, the communication costs of existing protocols mainly depend on the size of the larger set, leading to high communication cost. In this paper, we propose a low communication‐cost PSI protocol designed specifically for unbalanced two‐party private sets, aiming to enhance the efficiency of communication. For each item in the smaller set, the receiver queries whether it belongs to the larger set, such that the communication cost depends solely on the smaller set. The queries are implemented by private information retrieval which is constructed with trapdoor hash function. Our investigation indicates that in each instance of invoking the trapdoor hash function, the receiver is required to transmit both a hash key and an encoding key to the sender, thus incurring significant communication cost. In order to address this concern, we propose the utilization of a seed hash key, a seed encoding key, and a Latin square. By employing these components, the sender can autonomously generate all the necessary hash keys and encoding keys, obviating the multiple transmissions of such keys. The proposed protocol is provably secure against a semihonest adversary under the Decisional Diffie–Hellman assumption. Through implementation demonstration, we showcase that when the sizes of the two sets are 2 8 and 2 14 , the communication cost of our protocol is only 3.3% of the state‐of‐the‐art protocol and under 100 Kbps bandwidth, we achieve 1.46x speedup compared to the state‐of‐the‐art protocol. Our source code is available on GitHub: https://github.com/TAN-OpenLab/Unbanlanced-PSI .

ZNS SSDs divide the storage space into sequential-write zones, reducing costs of DRAM utilization, garbage collection, and over-provisioning. The sequential-write feature of zones is well-suited for LSM-based databases, where random writes are organized into sequential writes to improve performance. However, the current compaction mechanism of LSM-tree results in widely varying access frequencies (i.e., hotness) of data and thus incurs an extreme imbalance in the distribution of erasure counts across zones. The imbalance significantly limits the lifetime of SSDs. Moreover, the current zone-reset method involves a large number of unnecessary erase operations on unused blocks, further shortening the SSD lifetime. Considering the access pattern of LSM-tree, this article proposes a wear-aware zone-management technique, termed WA-Zone , to effectively balance inter- and intra-zone wear in ZNS SSDs. In WA-Zone, a wear-aware zone allocator is first proposed to dynamically allocate data with different hotness to zones with corresponding lifetimes, enabling an even distribution of the erasure counts across zones. Then, a partial-erase-based zone-reset method is presented to avoid unnecessary erase operations. Furthermore, because the novel zone-reset method might lead to an unbalanced distribution of erasure counts across blocks in a zone, a wear-aware block allocator is proposed. Experimental results based on the FEMU emulator demonstrate the proposed WA-Zone enhances the ZNS-SSD lifetime by 5.23×, compared with the baseline scheme.

The NVMe Zoned Namespace (ZNS) is a high-performance interface for flash-based solid-state drives (SSDs), which divides the logical address space into fixed-size and sequential-write zones. Meanwhile, ZNS SSDs eliminate in-device garbage collection (GC) by shifting the responsibility of GC to the host. However, the host-side GC of ZNS SSDs is not efficient. On the one hand, data migration during GC first moves data to the host buffer and then writes back the transferred data to the new location in the SSD, resulting in an unnecessary end-to-end transfer overhead. On the other hand, due to the pre-configured mapping between zones and blocks, GC incurs a large block-to-block rewrite overhead, i.e., even if most of the data in a block of the victim zone is valid, the valid data will still be rewritten to another block in the target zone. To address these issues, this article proposes a novel ZNS SSD design that features dynamic zone mapping, termed Brick-ZNS . Brick-ZNS implements two key functionalities: in-storage data migration and address remapping. New ZNS commands are first designed to realize in-storage data migration to avoid the end-to-end transfer overhead of GC while ensuring performance predictability. Then, a remapping strategy exploiting parallel physical blocks is proposed to reduce the large block-to-block rewrite overhead while ensuring zone-level access parallelism. The basic idea of the strategy is to directly remap the parallel physical blocks with a sufficient amount of valid data in the victim zone to the target zone, hence avoiding the large block-to-block rewrite overhead. Based on a full-stack SSD emulator, the evaluation results show that Brick-ZNS improves write throughput by 25% and SSD lifetime by 1.41×.

The NVMe Zoned Namespace (ZNS) is a new type of storage interface, which divides logical address space into fixed-size zones, and each zone strictly follows a sequential write constraint with a write pointer. Owing to the sequential write constraint of the ZNS, I/O requests would not be scheduled arbitrarily like the traditional SSDs with block interface. When multiple applications concurrently access one ZNS SSD hardware, the constraint deteriorates I/O blocking and causes huge unfairness. To resolve the problem, we propose a self-balance I/O scheduling dedicated for ZNS SSDs, called Fair-ZNS, to balance the slowdown among multiple applications and ensure fairness. Fair-ZNS identifies the unfair requests by the maximum slowdown value, and violently schedules these requests into the head of the queues overcoming the sequential write constraint. To eliminate the negative effect of the violent scheduling, Fair-ZNS deploys a self-balancing coordinator to fine-tune the order of the requests. Comprehensive evaluations show that Fair-ZNS alleviates I/O blocking and reduces average waiting time by 8.3×, increases fairness by 2.3×, and decreases the max slowdown by 5.1× averagely when compared to the current ZNS SSDs.

ZNS SSDs shift the responsibility of garbage collection (GC) to the host. However, data migration in GC needs to move data to the host's buffer first and write back to the new location, resulting in an unnecessary end-to-end transfer overhead. Moreover, due to the pre-configured mapping between zones and blocks, GC needs to perform a large number of unnecessary block-to-block data migrations between zones. To address these issues, this paper proposes a simple and efficient data migration method, called IS-AR, with in-storage data migration and address remapping. Based on a full-stack SSD emulator, our evaluation shows that IS-AR reduces GC latency by 6.78× and improves SSD lifetime by 1.17× on average.

Recently rumors have been rapidly propagated while the Internet has been extensively developed. Research shows that highly credible comments with a distinct stance have worthy information. In this paper, we attempt to combine user credibility and user stance to capture worthy comments during the information-dissemination process to detect rumors. We propose a User Stance Bi-Directional Graph Attention Networks (USBGAT) model to extract accurate information for rumor detection based on high credibility users with strong stance, and diminish ineffectively neutral comments. Specifically, we take user features and user stance as a component of the node features, with multiviews features of tweets content engaged. Then, we use bidirectional graph attention networks (GAT) to capture the high-level representation of the rumor. Furthermore, we reweight the node features according to users' stances. Extensive experiments on two datasets: Pheme and Weibo show that our model is superior to the state-of-the-art models, especially in the early rumor detection. Our code and data are available at https://github.com/TAN-OpenLab/USB-GAT

Lipreading is a task that involves analyzing and learning patterns and features of lip and mouth movements, extracting relevant information from videos, and converting it into understandable language or text. Correctly interpreting key frames in videos is crucial for the success of lipreading. In existing lipreading tasks, there has been limited research on the combination of temporal modeling with long-range dependencies and key frames, leading to the misclassification of phonetically similar words. In this work, we explore the integration of key frames into long-range dependency information and propose a joint perception temporal convolutional network (JP-TCN) for word-level lipreading. Specifically, a multi-scale feature extraction module is employed to extract local dynamic information and long-range dependency information from the front-end image features. Subsequently, these two sets of information are fed into an adaptive fusion module that weights and combines the local dynamic information and long-range dependency information at each time position, enabling the model to enhance key frames learning at different levels. Without any additional bells and whistles, the proposed JP-TCN achieves a recognition accuracy of 89.92% on the LRW dataset.

The research on detecting violent behavior in videos has made good progress, which provides good support for monitoring abnormal videos spread in the network, so as to achieve the effect of purifying the network space environment. A large number of current violence detection models have achieved good performance in experimental environments, but their generalization ability is insufficient. Violent behavior often occurs in a variety of scenarios, automatic detection of violent behavior requires a model with strong generalization. In this paper, a crowd violence behavior detection model with good generalization ability based on human contour and dynamic characteristics was designed. The model generalization ability is improved by focusing on the human features in the video and using the human dynamic features obtained from adjacent frames. In our model, a 3D-CNN framework was used to extract spatial features of the input feature map, and LSTM was used to fuse the temporal feature, we call this model HD-Net. Through multiple contrast experiments, the generalization ability of HD-Net is tested on three datasets: RLVS, Hockey and violent flow. Comparing with other classical violence detection models, the good generalization ability of the model is verified.

In recent multi-speaker speech separation researches, the overall deep-learning-based architecture consists of three parts: encoder, separator, and decoder. But improvement strategies generally only focus on the separator in the middle, regardless of its input. The most common encoder structure at present is a single 1D convolution layer followed by a nonlinear activation function, ReLU. In this paper, we firstly propose a new encoder named Attention DE, trying to improve the input effectiveness of the separator. The new encoder adds extra 1D convolutional layers and the multi-head attention mechanism to enhance the feature aggregation ability of input speech. Secondly, instead of RNNs, our separator uses SepFormer Blocks to improve the training efficiency and learn the speech sequence patterns better. Experiments show that the Attention DE is generally applicable to improve the performance of the single-channel speech separation model based on the time domain. The method of Attention DE fusion SepFormer blocks achieves an advanced SI-SNRi of 20.3dB on WSJ0-2MIX. Code is publicly available at https://github.com/TAN-OpenLab/AttentionDE.

Modern NVMe SSDs have been widely deployed in multi-tenant cloud computing environments or multi-programming systems. When multiple applications concurrently access one SSD hardware, unfairness within the shared SSD will slow down the application significantly and lead to a violation of service level objectives. However, traditional data cache management within SSDs mainly focuses on improving cache hit ratio, which causes data cache contention and sacrifices fairness among multiple applications. In this paper, we propose a DRAM-based Over-Provisioning (OP) cache management mechanism, named Justitia, to reduce data cache contention and improve fairness for modern SSDs. Justitia consists of two stages includingStatic-OPstage andDynamic-OPstage. Through the novel OP mechanism in the two stages, Justitia reduces the max slowdown by$4.5\times$on average. At the same time, Justitia increases fairness by$20.6\times$and buffer hit ratio by$19.6\%$averagely, compared with the traditional shared mechanism.

Social networks have been part of human beings' daily lives and affect nearly every aspect of our lives. Social influence prediction is an interesting topic to predict whether users will or will not be activated by current social spreading events, and deep learning-based approaches can obtain outstanding accuracy by graph neural networks (GNNs). However, GNN models are restricted by the 1-Weisfeiler-Lehman (WL) test and represent the node structure by only the first layer neighbors but cannot discriminate nodes with different second or more layer neighbors. Therefore, we propose a multilayer relation attention-based social influence prediction Net using GAT with local stimulation, named after MRAInf. Specifically, we design an enhanced node representation (ENR) to describe the original node structure vector by three-layer-neighbor adjacency relations, with more details for attention in GAT. Moreover, in GAT, we design a local stimulation (LS) mechanism with multiple 1D convolutions to reinforce the feature map of the target node being predicted and to weaken the feature maps of non-target nodes. Detailed latent information on ENR and local stimulation for targets in GAT benefit social influence prediction. We conduct extensive experiments on three benchmark datasets: Twitter, Open Academic Graph, and Digg, and the experimental results show that our approach outperforms existing comparison methods in terms of classification performance and predictive accuracy.

With the development and innovation of new techniques for 5G, 5G networks can provide extremely large capacity, robust integrity, high bandwidth, and low latency for multimedia image sharing and storage. However, it will surely exacerbate the privacy problems intrinsic to image transformation. Due to the high security and reliability requirements for storing and sharing sensitive images in the 5G network environment, verifiable steganography-based secret image sharing (SIS) is attracting increasing attention. The verifiable capability is necessary to ensure the correct image reconstruction. From the literature, efficient cheating verification, lossless reconstruction, low reconstruct complexity, and high-quality stego images without pixel expansion are summarized as the primary goals of proposing an effective steganography-based SIS scheme. Compared with the traditional underlying techniques for SIS, cellular automata (CA) and matrix projection have more strengths as well as some weaknesses. In this paper, we perform a complimentary of these two techniques to propose a verifiable secret image sharing scheme, where CA is used to enhance the security of the secret image, and matrix projection is used to generate shadows with a smaller size. From the steganography perspective, instead of the traditional least significant bits replacement method, matrix encoding is used in this paper to improve the embedding efficiency and stego image quality. Therefore, we can simultaneously achieve the above goals and achieve proactive and dynamic features based on matrix projection. Such features can make the proposed SIS scheme more applicable to flexible 5G networks. Finally, the security analysis illustrates that our scheme can effectively resist the collusion attack and detect the shadow tampering over the persistent adversary. The analyses for performance and comparative demonstrate that our scheme is a better performer among the recent schemes with the perspective of functionality, visual quality, embedding ratio, and computational efficiency. Therefore, our scheme further strengthens security for the images in 5G networks.

This paper analyzes the limitation of baseline estimation by defining four kinds of rating personalization corresponding to four kinds of users' rating criterions, including Normal, Strict, Lenient, and Middle. We find a standard deviation proportion pattern from ratings' normal distribution to enhance the handling capability of users' personalized rating behavior, and propose a novel baseline estimation model based on Standard Deviation Proportion, named SDP model, to improve the accuracy of existing recommendation algorithms which used traditional baseline estimation. We also propose two application instances of SDP, including SDPSVD++ and SDPTrustSVD, to show how to apply the proposed SDP. Experiments show that the SDP can not only improve the baseline estimation performance, but also can effectively improve predictive accuracies of existing recommendation algorithms.

In this paper, we propose a decentralized identity management system based on Blockchain. The function of the system mainly includes identity authentication and reputation management. The technical advantages of the Blockchain makes the data in the system safe and credible. In addition, we use smart contracts to write system rules to ensure the reliability of user information. We bind the user's entity information with the public key address and determine the true identity of a virtual user on the Blockchain. We use the token to represent the reputation which is shown to be an effective reputation model, making the participants in the system prefer to maintain and manage their personal reputation. Our system makes it possible for users to securely manage their identity and reputation on the Internet.

In recent years, trust has emerged with the development of cloud computing. It is a critical step to select a trusted cloud provider before the service begins, which is related to the interests of cloud consumers themselves and the quality of the service. A SLA trust model based on behavior evaluation is proposed in this paper. User's subjective evaluations are abandoned, the provider who is trusted and meets the demand for cloud consumers is selected according to the transaction history (mainly the parameter vectors formed during the transaction process) between cloud providers and cloud consumers and trust value of cloud providers before the service starts. In the service process, using iterative methods dynamically updates the trust value based on the fulfillment of SLA parameters. At the same time, the time factor is taken into account, so that the trust value is more reasonable. Experiments show that the model is able to select trusted provider to trade according to the demands of cloud consumers, dynamically updates the trust value and the trust value of malicious providers can be suppressed.